Comment on “A Bayesian Sensitivity Analysis of Out-of-Hospital 12-Lead Electrocardiograms: Implications for Regionalization of Cardiac Care”


To the Editor

I read with interest the article by Youngquist et al.1 suggesting that the low pretest probability of ST-elevation myocardial infarction (STEMI) in out-of-hospital patients with ischemic symptoms, combined with an imperfect electrocardiogram (ECG) specificity for STEMI, will unacceptably lower the positive predictive value of an out-of-hospital ECG. In April 2005 when we set up our system for out-of-hospital activation, without any base physician involvement, we required both a computer-aided ECG diagnosis of **Acute MI** and the presence of chest pain. Dyspnea, cardiac arrest, or any other presentation that did not include chest pain was not adequate for the paramedic decision to activate the catheterization laboratory. Through July 2007, we had 55 patients with STEMI correctly activated by the emergency medical services (EMS) and only four false-positive activations, for a positive predictive value of 93%. Since January 2004, a total of 68 of 104 (65%) STEMI patients transported by EMS would have been eligible for activation based on these criteria. The mean door-to-balloon time with out-of-hospital activation decreased from 86 to 56 minutes for those who have chest pain and **Acute MI** by the algorithm.2